An efficient GPU-based ALE-DEM coupling method for tire-sand interaction dynamics

The coupling of the finite element method and the discrete element method is one of the powerful approaches for simulating tire-sand interaction dynamics. However, the computational cost is significantly heavy for the models with a huge number of sand particles and fine meshes of the tire for the accurate contact detection. In this study, by introducing the arbitrary Lagrange-Euler (ALE) formulation, an efficient GPU-based ALE-DEM coupling method is proposed for simulating tire-sand interaction dynamics. Firstly, based on the ALE method and absolute nodal coordinate formulation (ANCF), the tire is discretized by the ALE-ANCF shell elements, and an incompressibility Mooney-Rivlin constitutive model is employed. A kinematic constraint equation with a material flow velocity is established to describe the relationship between the rigid rim and the flexible tire tread. By utilizing the ALE method, fine meshes of the tire tread are used on the contact region to obtain a detailed distribution of contact stress and strain, while coarse meshes are used on the other non-contact region to save computational time. Secondly, the sand particles are modeled by the DEM. Thirdly, for contact detection between finite elements and particles, an efficient global search algorithm based on the ALE method and an improved local search algorithm are proposed. Finally, the computational framework of the ALE-DEM coupling method based on GPU parallelization is established for the tire-sand system. Three numerical examples are presented to validate this method and the tire-sand interaction dynamics is simulated and comparatively studied.